Presentation Slides - Kosuke Imai

Transcription

Unpacking the Black Box of Causality:
Learning about Causal Mechanisms from
Experimental and Observational Studies
Kosuke Imai
Princeton University
Talk at the Center for Lifespan Psychology
Max Planck Institute for Human Development
April 28, 2015
Joint work with
Luke Keele Dustin Tingley Teppei Yamamoto
Project References (click the article titles)
General:
Unpacking the Black Box of Causality: Learning about Causal
Mechanisms from Experimental and Observational Studies.
American Political Science Review
Theory:
Identification, Inference, and Sensitivity Analysis for Causal
Mediation Effects. Statistical Science
Extensions:
A General Approach to Causal Mediation Analysis. Psychological
Methods
Experimental Designs for Identifying Causal Mechanisms. Journal
of the Royal Statistical Society, Series A (with discussions)
Identification and Sensitivity Analysis for Multiple Causal
Mechanisms: Revisiting Evidence from Framing Experiments.
Political Analysis
Software:
mediation: R Package for Causal Mediation Analysis. Journal of
Statistical Software
Kosuke Imai (Princeton)
Causal Mechanisms
Max Planck (April, 2015)
2 / 49
Identification of Causal Mechanisms
Causal inference is a central goal of scientific research
Scientists care about causal mechanisms, not just about causal
effects
Randomized experiments often only determine whether the
treatment causes changes in the outcome
Not how and why the treatment affects the outcome
Common criticism of experiments and statistics:
black box view of causality
Question: How can we learn about causal mechanisms from
experimental and observational studies?
Causal Mechanisms
3 / 49
Overview of the Talk
Present a general framework for statistical analysis and research
design strategies to understand causal mechanisms
1
Show that the sequential ignorability assumption is required to
identify mechanisms even in experiments
2
Offer a flexible estimation strategy under this assumption
3
Introduce a sensitivity analysis to probe this assumption
4
Illustrate how to use statistical software mediation
5
Consider research designs that relax sequential ignorability
Causal Mechanisms
4 / 49
Causal Mediation Analysis
Graphical representation
Mediator, M
Treatment, T
Outcome, Y
Goal is to decompose total effect into direct and indirect effects
Alternative approach: decompose the treatment into different
components
Causal mediation analysis as quantitative process tracing
Causal Mechanisms
5 / 49
Decomposition of Incumbency Advantage
Incumbency effects: one of the most studied topics in American
politics
Consensus emerged in 1980s: incumbency advantage is positive
and growing in magnitude
New direction in 1990s: Where does incumbency advantage
come from?
Scare-off/quality effect (Cox and Katz): the ability of incumbents to
deter high-quality challengers from entering the race
Alternative causal mechanisms: name recognition, campaign
spending, personal vote, television, etc.
Causal Mechanisms
6 / 49
Causal Mediation Analysis in Cox and Katz
Quality of challenger, M
Incumbency, T
Electoral outcome, Y
How much of incumbency advantage can be explained by
scare-off/quality effect?
How large is the mediation effect relative to the total effect?
Causal Mechanisms
7 / 49
Psychological Study of Media Effects
Large literature on how media influences public opinion
A media framing experiment of Brader et al.:
1
(White) Subjects read a mock news story about immigration:
Treatment: Hispanic immigrant in the story
Control: European immigrant in the story
2
Measure attitudinal and behavioral outcome variables:
Opinions about increasing or decrease immigration
Contact legislator about the issue
Send anti-immigration message to legislator
Why is group-based media framing effective?: role of emotion
Hypothesis: Hispanic immigrant increases anxiety, leading to
greater opposition to immigration
The primary goal is to examine how, not whether, media framing
shapes public opinion
Causal Mechanisms
8 / 49
Causal Mediation Analysis in Brader et al.
Anxiety, M
Media Cue, T
Immigration Attitudes, Y
Does the media framing shape public opinion by making people
anxious?
An alternative causal mechanism: change in beliefs
Can we identify mediation effects from randomized experiments?
Causal Mechanisms
9 / 49
The Standard Estimation Method
Linear models for mediator and outcome:
Yi = α1 + β1 Ti + ξ1> Xi + 1i
Mi = α2 + β2 Ti + ξ2> Xi + 2i
Yi = α3 + β3 Ti + γMi + ξ3> Xi + 3i
where Xi is a set of pre-treatment or control variables
1
2
3
4
Total effect (ATE) is β1
Direct effect is β3
Indirect or mediation effect is β2 γ
Effect decomposition: β1 = β3 + β2 γ.
Some motivating questions:
1
2
3
4
What should we do when we have interaction or nonlinear terms?
What about other models such as logit?
In general, under what conditions can we interpret β1 and β2 γ as
causal effects?
What do we really mean by causal mediation effect anyway?
Causal Mechanisms
10 / 49
Potential Outcomes Framework of Causal Inference
Observed data:
Binary treatment: Ti ∈ {0, 1}
Mediator: Mi ∈ M
Outcome: Yi ∈ Y
Observed pre-treatment covariates: Xi ∈ X
Potential outcomes model (Neyman, Rubin):
Potential mediators: Mi (t) where Mi = Mi (Ti )
Potential outcomes: Yi (t, m) where Yi = Yi (Ti , Mi (Ti ))
Total causal effect:
τi ≡ Yi (1, Mi (1)) − Yi (0, Mi (0))
Fundamental problem of causal inference: only one potential
outcome can be observed for each i
Causal Mechanisms
11 / 49
Back to the Examples
Mi (1):
1
2
Quality of her challenger if politician i is an incumbent
Level of anxiety individual i would report if he reads the story with
Hispanic immigrant
Yi (1, Mi (1)):
1
2
Election outcome that would result if politician i is an incumbent
and faces a challenger whose quality is Mi (1)
Immigration attitude individual i would report if he reads the story
with Hispanic immigrant and reports the anxiety level Mi (1)
Mi (0) and Yi (0, Mi (0)) are the converse
Causal Mechanisms
12 / 49
Causal Mediation Effects
Causal mediation (Indirect) effects:
δi (t) ≡ Yi (t, Mi (1)) − Yi (t, Mi (0))
Causal effect of the change in Mi on Yi that would be induced by
treatment
Change the mediator from Mi (0) to Mi (1) while holding the
treatment constant at t
Represents the mechanism through Mi
Zero treatment effect on mediator =⇒ Zero mediation effect
Examples:
1
2
Part of incumbency advantage that is due to the difference in
challenger quality induced by incumbency status
Difference in immigration attitudes that is due to the change in
anxiety induced by the treatment news story
Causal Mechanisms
13 / 49
Total Effect = Indirect Effect + Direct Effect
Direct effects:
ζi (t) ≡ Yi (1, Mi (t)) − Yi (0, Mi (t))
Causal effect of Ti on Yi , holding mediator constant at its potential
value that would realize when Ti = t
Change the treatment from 0 to 1 while holding the mediator
constant at Mi (t)
Represents all mechanisms other than through Mi
Total effect = mediation (indirect) effect + direct effect:
τi = δi (t) + ζi (1 − t) =
1
{(δi (0) + ζi (0)) + (δi (1) + ζi (1))}
2
Causal Mechanisms
14 / 49
Mechanisms, Manipulations, and Interactions
Mechanisms
Indirect effects: δi (t) ≡ Yi (t, Mi (1)) − Yi (t, Mi (0))
Counterfactuals about treatment-induced mediator values
Manipulations
Controlled direct effects: ξi (t, m, m0 ) ≡ Yi (t, m) − Yi (t, m0 )
Causal effect of directly manipulating the mediator under Ti = t
Interactions
Interaction effects: ξ(1, m, m0 ) − ξ(0, m, m0 )
The extent to which controlled direct effects vary by the treatment
Causal Mechanisms
15 / 49
What Does the Observed Data Tell Us?
Recall the Brader et al. experimental design:
1
2
randomize Ti
measure Mi and then Yi
Among observations with Ti = t, we observe Yi (t, Mi (t)) but not
Yi (t, Mi (1 − t)) unless Mi (t) = Mi (1 − t)
But we want to estimate
δi (t) ≡ Yi (t, Mi (1)) − Yi (t, Mi (0))
For t = 1, we observe Yi (1, Mi (1)) but not Yi (1, Mi (0))
Similarly, for t = 0, we observe Yi (0, Mi (0)) but not Yi (0, Mi (1))
We have the identification problem =⇒ Need assumptions or
better research designs
Causal Mechanisms
16 / 49
Counterfactuals in the Examples
1
Incumbency advantage:
An incumbent (Ti = 1) faces a challenger with quality Mi (1)
We observe the electoral outcome Yi = Yi (1, Mi (1))
We also want Yi (1, Mi (0)) where Mi (0) is the quality of challenger
this incumbent politician would face if she is not an incumbent
2
Media framing effects:
A subject viewed the news story with Hispanic immigrant (Ti = 1)
For this person, Yi (1, Mi (1)) is the observed immigration opinion
Yi (1, Mi (0)) is his immigration opinion in the counterfactual world
where he still views the story with Hispanic immigrant but his
anxiety is at the same level as if he viewed the control news story
In both cases, we can’t observe Yi (1, Mi (0)) because Mi (0) is not
realized when Ti = 1
Causal Mechanisms
17 / 49
Sequential Ignorability Assumption
Proposed identification assumption: Sequential Ignorability (SI)
{Yi (t 0 , m), Mi (t)} ⊥
⊥ Ti | Xi = x,
(1)
Yi (t 0 , m) ⊥
⊥ Mi (t) | Ti = t, Xi = x
(2)
In words,
1
2
Ti is (as-if) randomized conditional on Xi = x
Mi (t) is (as-if) randomized conditional on Xi = x and Ti = t
Important limitations:
1
2
3
4
In a standard experiment, (1) holds but (2) may not
Xi needs to include all confounders
Xi must be pre-treatment confounders =⇒ post-treatment
confounder is not allowed
Randomizing Mi via manipulation is not the same as assuming
Mi (t) is as-if randomized
Causal Mechanisms
18 / 49
Sequential Ignorability in the Standard Experiment
Back to Brader et al.:
Treatment is randomized =⇒ (1) is satisfied
But (2) may not hold:
1
2
Pre-treatment confounder or Xi : state of residence
those who live in AZ tend to have higher levels of perceived harm
and be opposed to immigration
Post-treatment confounder: alternative mechanism
beliefs about the likely negative impact of immigration makes
people anxious
Pre-treatment confounders =⇒ measure and adjust for them
Post-treatment confounders =⇒ adjusting is not sufficient
Causal Mechanisms
19 / 49
Nonparametric Identification
Under SI, both ACME and average direct effects are nonparametrically
identified (can be consistently estimated without modeling assumption)
¯
ACME δ(t)
Z Z
E(Yi | Mi , Ti = t, Xi ) {dP(Mi | Ti = 1, Xi ) − dP(Mi | Ti = 0, Xi )} dP(Xi )
¯
Average direct effects ζ(t)
Z Z
{E(Yi | Mi , Ti = 1, Xi ) − E(Yi | Mi , Ti = 0, Xi )} dP(Mi | Ti = t, Xi ) dP(Xi )
Implies the general mediation formula under any statistical model
Causal Mechanisms
20 / 49
Traditional Estimation Methods: LSEM
Linear structural equation model (LSEM):
Mi
= α2 + β2 Ti + ξ2> Xi + i2 ,
Yi
= α3 + β3 Ti + γMi + ξ3> Xi + i3 .
Fit two least squares regressions separately
Use product of coefficients (βˆ2 γˆ ) to estimate ACME
Use asymptotic variance to test significance (Sobel test)
¯
¯
Under SI and the no-interaction assumption (δ(1)
6= δ(0)),
βˆ2 γˆ
consistently estimates ACME
Can be extended to LSEM with interaction terms
Problem: Only valid for the simplest LSEM
Causal Mechanisms
21 / 49
Popular Baron-Kenny Procedure
The procedure:
1
2
3
Regress Y on T and show a significant relationship
Regress M on T and show a significant relationship
Regress Y on M and T , and show a significant relationship
between Y and M
The problems:
1
2
3
First step can lead to false negatives especially if indirect and direct
effects in opposite directions
The procedure only anticipates simplest linear models
Don’t do star-gazing. Report quantities of interest
Causal Mechanisms
22 / 49
Proposed General Estimation Algorithm
1
Model outcome and mediator
Outcome model: p(Yi | Ti , Mi , Xi )
Mediator model: p(Mi | Ti , Xi )
These models can be of any form (linear or nonlinear, semi- or
nonparametric, with or without interactions)
2
Predict mediator for both treatment values (Mi (1), Mi (0))
3
Predict outcome by first setting Ti = 1 and Mi = Mi (0), and then
Ti = 1 and Mi = Mi (1)
4
Compute the average difference between two outcomes to obtain
a consistent estimate of ACME
5
Monte-Carlo or bootstrap to estimate uncertainty
Causal Mechanisms
23 / 49
Example: Binary Mediator and Outcome
Two logistic regression models:
Pr(Mi = 1 | Ti , Xi ) = logit−1 (α2 + β2 Ti + ξ2> Xi )
Pr(Yi = 1 | Ti , Mi , Xi ) = logit−1 (α3 + β3 Ti + γMi + ξ3> Xi )
Can’t multiply β2 by γ
Difference of coefficients β1 − β3 doesn’t work either
Pr(Yi = 1 | Ti , Xi ) = logit−1 (α1 + β1 Ti + ξ1> Xi )
Can use our algorithm (example: E{Yi (1, Mi (0))})
1
2
Predict Mi (0) given Ti = 0 using the first model
b i (0), Xi ) using the
Compute Pr(Yi (1, Mi (0)) = 1 | Ti = 1, Mi = M
second model
Causal Mechanisms
24 / 49
Sensitivity Analysis
Standard experiments require sequential ignorability to identify
mechanisms
The sequential ignorability assumption is often too strong
Need to assess the robustness of findings via sensitivity analysis
Question: How large a departure from the key assumption must
occur for the conclusions to no longer hold?
Parametric sensitivity analysis by assuming
{Yi (t 0 , m), Mi (t)} ⊥
⊥ Ti | Xi = x
but not
Yi (t 0 , m) ⊥
⊥ Mi (t) | Ti = t, Xi = x
Possible existence of unobserved pre-treatment confounder
Causal Mechanisms
25 / 49
Parametric Sensitivity Analysis
Sensitivity parameter: ρ ≡ Corr(i2 , i3 )
Sequential ignorability implies ρ = 0
Set ρ to different values and see how ACME changes
Result:
β 2 σ1
¯
¯
δ(0)
= δ(1)
=
σ2
q
2
2
ρ˜ − ρ (1 − ρ˜ )/(1 − ρ ) ,
where σj2 ≡ var(ij ) for j = 1, 2 and ρ˜ ≡ Corr(i1 , i2 ).
When do my results go away completely?
¯ = 0 if and only if ρ = ρ˜
δ(t)
Easy to estimate from the regression of Yi on Ti :
Yi = α1 + β1 Ti + i1
Causal Mechanisms
26 / 49
Interpreting Sensitivity Analysis with R squares
Interpreting ρ: how small is too small?
An unobserved (pre-treatment) confounder formulation:
i2 = λ2 Ui + 0i2
and
i3 = λ3 Ui + 0i3
How much does Ui have to explain for our results to go away?
Sensitivity parameters: R squares
1
Proportion of previously unexplained variance explained by Ui
2∗
RM
≡ 1−
2
var(0i2 )
var(i2 )
and RY2∗ ≡ 1 −
var(0i3 )
var(i3 )
Proportion of original variance explained by Ui
0
e 2 ≡ var(i2 ) − var(i2 )
R
M
var(Mi )
and
Causal Mechanisms
0
e 2 ≡ var(i3 ) − var(i3 )
R
Y
var(Yi )
27 / 49
2∗ , R 2∗ ) (or (R
e2 , R
e 2 )):
Then reparameterize ρ using (RM
Y
M
Y
eY
eM R
sgn(λ2 λ3 )R
∗ ∗
ρ = sgn(λ2 λ3 )RM
,
RY = q
2 )(1 − R 2 )
(1 − RM
Y
2 and R 2 are from the original mediator and outcome
where RM
Y
models
sgn(λ2 λ3 ) indicates the direction of the effects of Ui on Yi and Mi
2∗ , R 2∗ ) (or (R
e2 , R
e 2 )) to different values and see how
Set (RM
Y
M
Y
mediation effects change
Causal Mechanisms
28 / 49
Reanalysis: Estimates under Sequential Ignorability
Original method: Product of coefficients with the Sobel test
— Valid only when both models are linear w/o T –M interaction
(which they are not)
Our method: Calculate ACME using our general algorithm
Outcome variables
Decrease Immigration
¯
δ(1)
Support English Only Laws
¯
δ(1)
Request Anti-Immigration Information
¯
δ(1)
Send Anti-Immigration Message
¯
δ(1)
Product of
Coefficients
Average Causal
Mediation Effect (δ)
.347
[0.146, 0.548]
.204
[0.069, 0.339]
.277
[0.084, 0.469]
.276
[0.102, 0.450]
.105
[0.048, 0.170]
.074
[0.027, 0.132]
.029
[0.007, 0.063]
.086
[0.035, 0.144]
Causal Mechanisms
29 / 49
0.4
0.2
0.0
−0.2
−0.4
Average Mediation Effect: δ(1)
Reanalysis: Sensitivity Analysis w.r.t. ρ
−1.0
−0.5
0.0
0.5
1.0
Sensitivity Parameter: ρ
ACME > 0 as long as the error correlation is less than 0.39
(0.30 with 95% CI)
Causal Mechanisms
30 / 49
0.5
0.4
−0
.15
0.3
−0.
1
5
0.1
0.2
−0.0
0
0.0
Proportion of Total Variance in Y
Explained by Confounder
˜ 2 and R
˜2
Reanalysis: Sensitivity Analysis w.r.t. R
M
Y
0.05
0.0
0.1
0.2 0.3 0.4 0.5 0.6 0.7
Proportion of Total Variance in M
Explained by Confounder
0.8
An unobserved confounder can account for up to 26.5% of the variation
in both Yi and Mi before ACME becomes zero
Causal Mechanisms
31 / 49
Open-Source Software “Mediation”
Causal Mechanisms
32 / 49
Implementation Examples
1
Fit models for the mediator and outcome variable and store these
models
> m <- lm(Mediator ~ Treat + X)
> y <- lm(Y ~ Treat + Mediator + X)
2
Mediation analysis: Feed model objects into the mediate()
function. Call a summary of results
> m.out<-mediate(m, y, treat = "Treat",
mediator = "Mediator")
> summary(m.out)
3
Sensitivity analysis: Feed the output into the medsens() function.
Summarize and plot
>
>
>
>
s.out <- medsens(m.out)
summary(s.out)
plot(s.out, "rho")
plot(s.out, "R2")
Causal Mechanisms
33 / 49
Data Types Available via mediation
Outcome Model Types
Mediator Model Types
Linear
GLM
Ordered
Censored
Quantile
GAM
Survival
Linear (lm/lmer)
X
X
X∗
X
X
X∗
X
GLM (glm/bayesglm/glmer)
X
X
X∗
X
X
X∗
X
Ordered (polr/bayespolr)
X
X
X∗
X
X
X∗
X
-
-
-
-
-
-
-
Quantile (rq)
X∗
X∗
X∗
X∗
X∗
X∗
X
GAM (gam)
X∗
X∗
X∗
X∗
X∗
X∗
X∗
X
X
X∗
X
X
X∗
X
Censored (tobit via vglm)
Survival (survreg)
Types of Models That Can be Handled by mediate. Stars (∗ ) indicate the model
combinations that can only be estimated using the nonparametric bootstrap (i.e. with
boot = TRUE).
Causal Mechanisms
34 / 49
Additional Features
Treatment/mediator interactions, with formal statistical tests
Treatment/mediator/pre-treatment interactions and reporting of
quantities by pre-treatment values
Factoral, continuous treatment variables
Cluster standard errors/adjustable CI reporting/p-values
Support for multiple imputation
Multiple mediators
Multilevel mediation (NEW!)
Please read our vignette file here.
Causal Mechanisms
35 / 49
Causal Mediation Analysis in Stata
Based on the same algorithm
Hicks, R, Tingley D. 2011. Causal Mediation Analysis. Stata Journal.
11(4):609-615.
ssc install mediation
More limited coverage of models (just bc. of time though!)
Causal Mechanisms
36 / 49
Syntax: medeff
medeff (equation 1) (equation 2) [if] [in] [[weight]] ,
[sims(integer) seed(integer) vce(vcetype) Level(#)
interact(varname)] mediate(varname) treat(varname)
Where “equation 1” or “equation 2” are of the form (For equation 1, the
mediator equation):
probit M T x
or
regress M T x
Causal Mechanisms
37 / 49
Beyond Sequential Ignorability
Without sequential ignorability, standard experimental design
lacks identification power
Even the sign of ACME is not identified
Need to develop alternative experimental designs for more
credible inference
Possible when the mediator can be directly or indirectly
manipulated
All proposed designs preserve the ability to estimate the ACME
under the SI assumption
Trade-off: statistical power
These experimental designs can then be extended to natural
experiments in observational studies
Causal Mechanisms
38 / 49
Parallel Design
Must assume no direct effect of manipulation on outcome
More informative than standard single experiment
If we assume no T –M interaction, ACME is point identified
Causal Mechanisms
39 / 49
Why Do We Need No-Interaction Assumption?
Numerical Example:
Prop.
0.3
0.3
0.1
0.3
Mi (1)
1
0
0
1
Mi (0)
0
0
1
1
Yi (t, 1)
0
1
0
1
Yi (t, 0)
1
0
1
0
δi (t)
−1
0
1
0
¯ = −0.2
E(Mi (1) − Mi (0)) = E(Yi (t, 1) − Yi (t, 0)) = 0.2, but δ(t)
The Problem: Causal effect heterogeneity
T increases M only on average
M increases Y only on average
T − M interaction: Many of those who have a positive effect of T on
M have a negative effect of M on Y (first row)
A solution: sensitivity analysis (see Imai and Yamamoto, 2013)
Pitfall of “mechanism experiments” or “causal chain approach”
Causal Mechanisms
40 / 49
Encouragement Design
Direct manipulation of mediator is difficult in most situations
Use an instrumental variable approach:
Advantage: allows for unobserved confounder between M and Y
Key Assumptions:
1
2
Z is randomized or as-if random
No direct effect of Z on Y (a.k.a. exclusion restriction)
Causal Mechanisms
41 / 49
Example: Social Norm Experiment on Property Taxes
Lucia Del Carpio. “Are Neighbors Cheating?”
Treatment: informing average rate of compliance
Outcome: compliance rate obtained from administrative records
Large positive effect on compliance rate ≈ 20 percentage points
Mediators:
1
2
3
social norm (not measured; direct effect)
M1 : beliefs about compliance (measured)
M2 : beliefs about enforcement (measured)
Instruments:
1
2
Z1 : informing average rate of enforcement
Z2 : payment-reminder
Assumptions:
1
2
Z1 affects Y only through M1 and M2
Z2 affects Y only through M1
Results:
Average direct effect is estimated to be large
The author interprets this effect as the effect of social norm
Causal Mechanisms
42 / 49
Crossover Design
Recall ACME can be identified if we observe Yi (t 0 , Mi (t))
Get Mi (t), then switch Ti to t 0 while holding Mi = Mi (t)
Crossover design:
1
2
Round 1: Conduct a standard experiment
Round 2: Change the treatment to the opposite status but fix the
mediator to the value observed in the first round
Very powerful – identifies mediation effects for each subject
Must assume no carryover effect: Round 1 must not affect Round
2
Can be made plausible by design
Causal Mechanisms
43 / 49
Example: Labor Market Discrimination
E XAMPLE Bertrand & Mullainathan (2004, AER)
Treatment: Black vs. White names on CVs
Mediator: Perceived qualifications of applicants
Outcome: Callback from employers
Quantity of interest: Direct effects of (perceived) race
Would Jamal get a callback if his name were Greg but his
qualifications stayed the same?
Round 1: Send Jamal’s actual CV and record the outcome
Round 2: Send his CV as Greg and record the outcome
Assumption: their different names do not change the perceived
qualifications of applicants
Under this assumption, the direct effect can be interpreted as
blunt racial discrimination
Causal Mechanisms
44 / 49
Designing Observational Studies
Key difference between experimental and observational studies:
treatment assignment
Sequential ignorability:
1
2
Ignorability of treatment given covariates
Ignorability of mediator given treatment and covariates
Both (1) and (2) are suspect in observational studies
Statistical control: matching, propensity scores, etc.
Search for quasi-randomized treatments: “natural” experiments
How can we design observational studies?
Experiments can serve as templates for observational studies
Causal Mechanisms
45 / 49
Cross-over Design in Observational Studies
E XAMPLE Back to incumbency advantage
Use of cross-over design (Levitt and Wolfram)
1
2
1st Round: two non-incumbents in an open seat
2nd Round: same candidates with one being an incumbent
Assume challenger quality (mediator) stays the same
Estimation of direct effect is possible
Redistricting as natural experiments (Ansolabehere et al.)
1
2
1st Round: incumbent in the old part of the district
2nd Round: incumbent in the new part of the district
Challenger quality is the same but treatment is different
Estimation of direct effect is possible
Causal Mechanisms
46 / 49
Multiple Mediators
Quantity of interest = The average indirect effect with respect to M
W represents the alternative observed mediators
Left: Assumes independence between the two mechanisms
Right: Allows M to be affected by the other mediators W
Applied work often assumes the independence of mechanisms
Under this independence assumption, one can apply the same
analysis as in the single mediator case
For causally dependent mediators, we must deal with the
heterogeneity in the T × M interaction as done under the parallel
design =⇒ sensitivity analysis
Causal Mechanisms
47 / 49
Concluding Remarks
Even in a randomized experiment, a strong assumption is needed
to identify causal mechanisms
However, progress can be made toward this fundamental goal of
scientific research with modern statistical tools
A general, flexible estimation method is available once we assume
sequential ignorability
Sequential ignorability can be probed via sensitivity analysis
More credible inferences are possible using clever experimental
designs
Insights from new experimental designs can be directly applied
when designing observational studies
Multiple mediators require additional care when they are causally
dependent
Causal Mechanisms
48 / 49
The project website for papers and software:
http://imai.princeton.edu/projects/mechanisms.html
Email for questions and suggestions:
kimai@princeton.edu
Causal Mechanisms
49 / 49

Presentation Slides - Kosuke Imai

Transcription

Similar documents

Bridal Show - Mercer Oaks Catering

Zhang Haihui, "Mo Youzhi and His Song Yuan jiubenshu jingyanlu

JaHyun Kim Haboush, "The Texts of the Memoirs of Lady Hyegyòng

Document 6576546

NewGreekSongs C y c l e s i n T r...

Market Research Stage 6 Business Studies

BBQ RIB COOK-OFF BBQ RIB COOK-OFF

The Princeton Footnotes

ETS Princeton Campus map

Syllabus and Reading List